The Best Strategy to break into tech in 2023 is Open Source.
Announcing The Complete Guide To Open Source Course
- 100% free
- Practical Examples
- Find real projects
- Less than 1 hour
Comment "open" below and I will DM it to you
(must be following)
Brother you mentioned so many technologies, which makes me think you're not even sure the source of your slow speed.
90% can bet it has nothing to do with Drizzle.
How does one fix the speed issue with NextJS 16 APIs, Postgres & Drizzle ORM. Sometimes it just loads basic cached items for up to 4 seconds, no one would wait for content to load for that long.
@arok3nn It's a vision model, so it can pin point exactly the cards, and figure out where the card starts and ends.
That's why it's so good.
It knows exactly what's a card, a title, the image, and it can just crop and extract the image.
You can try it yourself. It's impressive.
I thought I was the only one that noticed.
It is scary good.
Testing with someone's design from twitter a few days ago.
100% nailed it. Even exported the images and picked the perfect color codes.
We've raised $65 billion in Series H funding at a $965 billion post-money valuation, led by @AltimeterCap, Dragoneer, @Greenoaks, and @sequoia.
This investment will help us advance our research and expand our capacity to meet growing demand for Claude.
Quick 100M token test with Xiaomi Mimo V2p5 Pro's new pricing.
Sooo good.
For reference if you bought a $6/m price you get 4 billion tokens.
If you are part of us poors you should test this for your work.
@crazibeat1 Oh it's not about that.
I'm just making a joke about how nigerian devs like to throw numbers about scale, when we truly only have basic to medium scale numbers in the entirety of Africa.
Helped a startup do this once. With a proxy it's quite easy to blur out information.
For example, agents cannot do a select on users table, or delivery_addresses table.
Or you can let them select, and then redact data in response.
Not a big problem to be honest. 90% of your data in DB is usually non sensitive.
Yes you can give your agents full access to your prod DB, but not the actual prod db.
Setup replication. For small databases like YC DB, replication should take a few minutes, then point the agent to the replica, not the real db.
Also setup a proxy layer between agent and replica.
So they never run any queries that affect or read sensitive customer tables.
One night I quietly gave our AI agent full access to YC's production database. It made the agent 10x more useful. That's what convinced me that trust-by-default is the only way to get the most out of agents.